Associação de Dietas com Câncer

Prévia do material em texto

<p>Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>https://doi.org/10.1186/s12916‑022‑02257‑9</p><p>RESEARCH ARTICLE</p><p>Association of meat, vegetarian, pescatarian</p><p>and fish‑poultry diets with risk of 19 cancer sites</p><p>and all cancer: findings from the UK Biobank</p><p>prospective cohort study and meta‑analysis</p><p>Solange Parra‑Soto1,2†, Danay Ahumada3†, Fanny Petermann‑Rocha1,2,4, Jirapitcha Boonpoor2,</p><p>Jose Lara Gallegos5, Jana Anderson1, Linda Sharp6, Fiona C. Malcomson7, Katherine M. Livingstone8,</p><p>John C. Mathers7, Jill P. Pell1†, Frederick K. Ho1† and Carlos Celis‑Morales1,2,9*†</p><p>Abstract</p><p>Background: The associations of cancer with types of diets, including vegetarian, fish, and poultry‑containing diets,</p><p>remain unclear. The aim of this study was, therefore, to investigate the association of type of diet with all cancers and</p><p>19 site‑specific incident cancers in a prospective cohort study and then in a meta‑analysis of published prospective</p><p>cohort studies.</p><p>Methods: A total of 409,110 participants from the UK Biobank study, recruited between 2006 and 2010, were</p><p>included. The outcomes were incidence of all cancers combined and 19 cancer sites. Associations between the types</p><p>of diets and cancer were investigated using Cox proportional hazards models. Previously published prospective</p><p>cohort studies were identified from four databases, and a meta‑analysis was conducted using random‑effects models.</p><p>Results: The mean follow‑up period was 10.6 years (IQR 10.0; 11.3). Compared with meat‑eaters, vegetarians (hazard</p><p>ratio (HR) 0.87 [95% CI: 0.79 to 0.96]) and pescatarians (HR 0.93 [95% CI: 0.87 to 1.00]) had lower overall cancer risk.</p><p>Vegetarians also had a lower risk of colorectal and prostate cancers compared with meat‑eaters. In the meta‑analysis,</p><p>vegetarians (Risk Ratio (RR): 0.90 [0.86 to 0.94]) and pescatarians (RR 0.91 [0.86; 0.96]) had lower risk of overall and colo‑</p><p>rectal cancer. No associations between the types of diets and prostate, breast, or lung cancers were found.</p><p>Conclusions: Compared with meat‑eaters, vegetarians and pescatarians had a lower risk of overall, colorectal, and</p><p>prostate cancer. When results were pooled in a meta‑analysis, the associations with overall and colorectal cancer</p><p>persisted, but the results relating to other specific cancer sites were inconclusive.</p><p>Keywords: Cancer, Diet, Vegetarians, Pescatarian</p><p>© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which</p><p>permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the</p><p>original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or</p><p>other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line</p><p>to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory</p><p>regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this</p><p>licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco</p><p>mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.</p><p>Introduction</p><p>Unhealthy diets have been associated with a higher risk</p><p>of several adverse health outcomes, including cancer [1,</p><p>2]. Over 11 million deaths were attributed to poor diet in</p><p>2017, among which more than 930,000 were attributed</p><p>to cancer, particularly, breast and colorectal cancer [1].</p><p>Although there is considerable evidence regarding the</p><p>associations between diet and cancer risk, most studies</p><p>Open Access</p><p>*Correspondence: Carlos.Celis@glasgow.ac.uk</p><p>†Solange Parra‑Soto and Danay Ahumada contributed equally to this</p><p>work and are joint first authors.</p><p>†Jill P Pell, Frederick K. Ho and Carlos Celis‑Morales contributed equally to</p><p>this work and are joint senior authors.</p><p>2 British Heart Foundation Cardiovascular Research Centre, Institute</p><p>of Cardiovascular and Medical Sciences, University of Glasgow,</p><p>Glasgow G12 8TA, UK</p><p>Full list of author information is available at the end of the article</p><p>http://creativecommons.org/licenses/by/4.0/</p><p>http://creativecommons.org/publicdomain/zero/1.0/</p><p>http://creativecommons.org/publicdomain/zero/1.0/</p><p>http://crossmark.crossref.org/dialog/?doi=10.1186/s12916-022-02257-9&domain=pdf</p><p>Page 2 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>have focused on single nutrients [3, 4] or food items [3,</p><p>5], which do not provide insights into how the widely</p><p>varying combinations of these foods within whole dietary</p><p>patterns impact risk.</p><p>What we eat and the type of dietary pattern we follow</p><p>are influenced by socioeconomic status, environmental</p><p>factors, and cultural and personal beliefs [1]. Recently,</p><p>there has been a growing concern about the impact of</p><p>food consumption on not just human health but also</p><p>planetary health. This has led to an increasing number of</p><p>people worldwide changing from diets that include meat</p><p>to other types of diet such as vegetarian, vegan, and pes-</p><p>catarian [6–9]. Recent estimates indicate that 4% of the</p><p>worldwide population are vegetarian, with almost 40%</p><p>reporting frequent consumption of vegetarian meals</p><p>[10]. Although there is increasing evidence regarding the</p><p>health benefits associated with these types of diets, most</p><p>studies have focused on outcomes such as all-cause mor-</p><p>tality and cardiovascular diseases, with limited and con-</p><p>flicting evidence for all cancers combined and specific</p><p>cancers [11–14]. One study conducted in 73,308 Sev-</p><p>enth-day Adventist men and women who were followed</p><p>up for 5.7 years reported a 12% lower risk of all-cause</p><p>mortality in vegetarians compared with non-vegetarians</p><p>[14]. Although no associations were observed for the</p><p>incidence of all cancer combined or mortality [14], veg-</p><p>etarians had a lower risk of cancers of the gastrointesti-</p><p>nal tract [15]. Another study that pooled data from two</p><p>prospective cohorts, covering 61,647 British men and</p><p>women who were followed for 14.9 years, reported that</p><p>fish-eaters had 38%, 34% and 12% lower risk of stomach,</p><p>colorectal, and all cancers, respectively [13], while veg-</p><p>etarians had 63% and 12% lower risk of stomach and all</p><p>cancers, respectively, compared with meat-eaters [13]. A</p><p>combined analysis of EPIC-Oxford and the Oxford Vege-</p><p>tarian Study, also from the UK, reported that vegetarians</p><p>(including vegans) had a lower risk of all cancers and can-</p><p>cers of the stomach, bladder, lymphatic and haematopoi-</p><p>etic system but a higher risk of cervical cancer compared</p><p>with meat-eaters [16]. In the UK Women’s Cohort Study,</p><p>a vegetarian diet was not associated with differences in</p><p>the risk of breast cancer [17]. The latest meta-analysis for</p><p>overall cancer, published in 2017, pooled only two pro-</p><p>spective cohort studies and showed an 8% reduction in</p><p>overall incident cancers associated with vegetarian diets</p><p>[18]. Another meta-analysis, conducted in 2017, included</p><p>nine studies (n = 686,629 participants) and reported no</p><p>associations between a vegetarian diet and the risk of</p><p>breast, colorectal or prostate cancers [19].</p><p>In summary, existing evidence shows that investiga-</p><p>tions of associations with dietary patterns have been</p><p>restricted to a limited number of cancer sites with equiv-</p><p>ocal findings. We addressed these research gaps using</p><p>data from the UK Biobank, a large prospective cohort</p><p>study, to investigate the associations between the type of</p><p>diet and all incident cancers combined and 19 site-spe-</p><p>cific cancers. In addition, we combined our findings with</p><p>those from past studies to provide an up-to-date meta-</p><p>analysis of prospective cohort studies.</p><p>Methods</p><p>UK Biobank</p><p>Between April 2007 and December 2010, the UK Biobank</p><p>recruited over 500,000 participants, aged 37 to 73 years</p><p>from the general population [20]. Participants attended</p><p>one of 22 assessment centres across England, Wales and</p><p>Scotland [21, 22], where they completed a touchscreen</p><p>questionnaire, had physical measurements taken and</p><p>provided biological samples, as described in detail else-</p><p>where [21, 22].</p><p>Outcomes</p><p>The primary outcome was incident cancer defined as</p><p>the first record of hospitalization for cancer or death</p><p>due to cancer, if no prior record of hospitalization. Date</p><p>and cause of death were obtained from death certificates</p><p>available up to 1 June 2020. Dates and causes of hospi-</p><p>tal admissions were obtained from the Health Episode</p><p>Statistics (England and Wales) and Scottish Morbidity</p><p>Records (Scotland). Follow-up for incident events was</p><p>censored on this date or the date of the event (cancer</p><p>diagnosis or death), whichever came first. The Interna-</p><p>tional Classification of Diseases, 10th revision (ICD-10)</p><p>was used to define the overall cancer (C00–C97, exclud-</p><p>ing non-melanoma skin cancer (C44)) and the following</p><p>19 cancers and four subgroups of colorectal cancer: head</p><p>and neck (C00–C14), oesophagus (C15), stomach (C16),</p><p>colorectal (C18, C19 and C20), colon (C18.0), colon prox-</p><p>imal (C18.0–18.4), colon distal (C18.5, C18.7), rectum</p><p>(C19–C20), pancreas (C25), lung (C33–34), malignant</p><p>melanoma (C43), breast in premenopausal and postmen-</p><p>opausal women (C50), uterus (C54–C55), ovary (C56),</p><p>prostate (C61), kidney (C64–C65), bladder (C67), brain</p><p>(C70–72), lymphatic and haematopoietic tissue (C81–</p><p>C96), non-Hodgkin lymphoma (C82–C86, C96), multiple</p><p>myeloma (C88–C90) and leukaemia (C91–C95). Further</p><p>details of these measurements can be found in the UK</p><p>Biobank online protocol [23].</p><p>Types of diets</p><p>A touchscreen dietary intake questionnaire, containing</p><p>29 questions about diet and 18 questions about alcohol,</p><p>completed at recruitment (baseline), was used to collect</p><p>data on the frequency of food intake over the past year.</p><p>Participants chose a frequency of intake ranging from</p><p>“never” to “once or more daily” for each food item. The</p><p>Page 3 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>food items included were cheese, milk, fish (oily and non-</p><p>oily), poultry and red meat (processed meat, beef, lamb</p><p>or mutton, pork, chicken, turkey or other poultry). Par-</p><p>ticipants were also asked to report whether they followed</p><p>any particular diet, including gluten-free, lactose-free,</p><p>low calorie, vegetarian and vegan diets. Based on their</p><p>responses, participants were categorized into one of the</p><p>following diets: vegetarian, which included lacto-ovo-</p><p>vegetarian (who consumed cheese and/or milk but they</p><p>never consumed fish, poultry or red meat) and vegan</p><p>(who reported never consuming milk, cheese, fish, poul-</p><p>try or red meat); pescatarian (who consumed cheese,</p><p>milk and fish but never consumed poultry or red meat);</p><p>fish-poultry eaters (who consumed cheese, milk, fish and</p><p>poultry but never consumed red meat); and meat-eaters</p><p>(who consumed cheese, milk, fish, poultry and red meat).</p><p>Due to the low number of participants following a vegan</p><p>diet (n = 57), these were pooled with vegetarians. To take</p><p>account of people changing their dietary patterns, we</p><p>excluded people who self-reported at baseline that their</p><p>diet often varied (n = 45,028, 8.99%). In addition, those</p><p>participants who reported being vegetarians but who</p><p>self-reported eating any meat products were excluded</p><p>from the study (n = 57). The same approach was used for</p><p>pescatarians and fish-poultry eaters [24]. Dietary infor-</p><p>mation for total energy and macro- and micro-nutrients</p><p>was collected via the Oxford WebQ, a web-based 24-hour</p><p>dietary questionnaire [25]. Bradbury et al. reported that</p><p>data collected using the dietary touchscreen question-</p><p>naire, which was applied to the entire cohort, correctly</p><p>ranked subjects according to their primary food group</p><p>intakes [26].</p><p>Covariates</p><p>Sociodemographic factors (sex and ethnicity) were self-</p><p>reported at the baseline assessment visit using a touch-</p><p>screen questionnaire. Age was calculated from the date</p><p>of birth at baseline assessment. Area-based socioeco-</p><p>nomic status was derived from the postcode of residence</p><p>using the Townsend score (16), which generates a depri-</p><p>vation score based on four census variables: unemploy-</p><p>ment, non-car ownership, non-house ownership and</p><p>household overcrowding. Self-reported smoking sta-</p><p>tus was categorized as never, former or current smoker.</p><p>Body mass index (BMI) was calculated from weight and</p><p>height expressed in kg/m2, and the World Health Organi-</p><p>zation (WHO) criteria were applied to classify partici-</p><p>pants into underweight (</p><p>breast, prostate, digestive tract</p><p>and lung) cancers. Only prospective cohort studies were</p><p>included. As recorded in PROSPERO, two authors (SP-S</p><p>and DA) searched MEDLINE/PubMed, Scopus, Embase</p><p>and Web of Science using the search terms described in</p><p>the Additional file  1: Methods. Two stages of screening</p><p>(1) the title and abstract and (2) the full text of potentially</p><p>eligible papers were performed. Data extraction was car-</p><p>ried out independently by the authors SP-S and DA in</p><p>Rayyan, and the results were then extracted to an Excel</p><p>spreadsheet. The inclusion was restricted to prospective</p><p>cohort studies published any time up to and including 31</p><p>August 2021, which were conducted in adults, included</p><p>some/all of the following types of diets (meat, vegetar-</p><p>ian, pescatarian, fish and poultry diet), provided results</p><p>for some/all of the relevant cancer outcomes and was</p><p>written in English. We excluded case-control studies and</p><p>studies that did not define the type of diet. Meta-analysis</p><p>was undertaken using a random-effects model, strati-</p><p>fied by type of diet (vegetarian, pescatarian or both) and</p><p>only included specific cancer sites that were reported by</p><p>at least three independent studies. Manuscripts that met</p><p>the inclusion criteria were assessed independently by the</p><p>two authors using the RAYYAN software [29]. We used</p><p>funnel plots to assess the potential bias within the stud-</p><p>ies included in the meta-analyses. The quality of the stud-</p><p>ies was also assessed using the Newcastle-Ottawa Scale</p><p>(Additional file 1: Table S1) [30]. Heterogeneity between</p><p>studies was tested using the I2 statistic. All analyses were</p><p>undertaken using the R statistical software, version 3.6.2</p><p>with the package “meta”.</p><p>Results</p><p>UK Biobank</p><p>Of the 502,492 UK participants, 99,382 were excluded as</p><p>they had cancer diagnoses at baseline (n = 41,406) and</p><p>missing data for types of diet or relevant covariates (n</p><p>= 57,976) (baseline characteristics of people with miss-</p><p>ing variables are in Additional file 1: Table S2). Therefore,</p><p>this study included 409,110 participants, of whom 53.4%</p><p>were women. Overall, 7256 (1.8%) were vegetarian, 9498</p><p>(2.3%) were pescatarian, 4625 (1.1%) were fish-poultry</p><p>eaters and 387,731 (94.8%) were meat-eaters (Additional</p><p>file  1: Fig. S1). The median follow-up period was 10.6</p><p>years (interquartile range 10.0 to 11.3), and 39,596 par-</p><p>ticipants developed cancers during follow-up. The soci-</p><p>odemographic characteristics of the population by type</p><p>of diet are presented in Table 1. In comparison with the</p><p>other groups, meat-eaters were older and more likely</p><p>to be overweight or obese and to smoke (Table  1). The</p><p>characteristics of dietary intake across types of diets are</p><p>presented in Additional file 1: Table S2. Energy consump-</p><p>tion was similar in the diet groups. However, intake of</p><p>fibre, polyunsaturated fat and water was slightly higher</p><p>in participants with vegetarian and pescatarian diets. In</p><p>contrast, protein intake was lower in vegetarians (mean</p><p>12.4 ± 2.3 SD % of total energy) than in meat-eaters</p><p>(mean 15.7 ± 3.6 SD % of total energy); crisps and pizza</p><p>were more consumed by vegetarian and pescatarian.</p><p>Additional file  1: Table  S3 shows the characteristics of</p><p>the UK Biobank population excluded from the study (n</p><p>= 57,976). Briefly, compared with the cohort included in</p><p>the present study, those excluded were more likely to be</p><p>women, individuals from more deprived backgrounds, of</p><p>non-white ethnicity and have a higher BMI.</p><p>The findings for the associations of types of diets with</p><p>incident cancer are shown in Table 2. In the unadjusted</p><p>model, vegetarians had a lower risk of liver, pancre-</p><p>atic, lung, prostate, bladder, colorectal, melanoma, kid-</p><p>ney, non-Hodgkin lymphoma and lymphatic cancer as</p><p>well as overall cancer, with hazard ratios ranging from</p><p>0.29 to 0.70 (Table  2). However, when the models were</p><p>fully adjusted for sociodemographic and lifestyle fac-</p><p>tors, multimorbidity and BMI (model 4), the associa-</p><p>tions remained statistically significant only for prostate</p><p>cancer (HR 0.57 [95% CI 0.43 to 0.76]), colorectal can-</p><p>cer (HR 0.73 [95% CI 0.54; 0.99]) and all cancers com-</p><p>bined (HR 0.87 [95% CI 0.79 to 0.96]). When colorectal</p><p>cancer was stratified according to subtypes, a lower risk</p><p>was observed for colon (HR 0.69 [95% CI 0.48; 0.99]) and</p><p>proximal colon (HR 0.43 [95% CI 0.22; 0.82]) in vegetar-</p><p>ians compared with meat-eaters, but not for rectal or dis-</p><p>tal cancer.</p><p>Lower risk of overall cancer and nine cancer sites was</p><p>found for pescatarians compared with meat-eaters in</p><p>the unadjusted analyses—kidney, lung, melanoma, non-</p><p>Hodgkin lymphoma, colorectal (overall and for colon</p><p>and rectum individually), bladder, prostate, lymphatic</p><p>and breast—with hazard ratios ranging from 0.57 to 0.65.</p><p>However, in the maximally adjusted model (model 4),</p><p>only overall cancer (HR 0.93 [95% CI 0.87 to 1.00]) and</p><p>melanoma (HR 0.68 [95% CI 0.47; 0.98]) remained signif-</p><p>icant. The hazard ratios for intermediate models 1, 2 and</p><p>3 are presented in Additional file 1: Table S4.</p><p>Similar results were found when the models were</p><p>repeated using the 2-year landmark analysis. In the maxi-</p><p>mally adjusted models (model 4), the associations for</p><p>vegetarians were slightly attenuated but remained signifi-</p><p>cant for overall (HR 0.88 [95% CI 0.80 to 0.97]), proximal</p><p>colon (HR 0.48 [95% CI 0.25; 0.93]) and prostate (HR 0.59</p><p>[95% CI 0.44; 0.79]) cancer. However, the associations of</p><p>vegetarian diet with colorectal and colon cancer were no</p><p>Page 5 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>Table 1 Sociodemographic characteristics of the study population by types of diet</p><p>Multimorbidity was defined as the existence of 2 or more chronic diseases</p><p>BMI Body mass index, n Number, PA Physical activity, MET Metabolic equivalent, SD Standard deviation</p><p>Characteristics Meat‑eaters Vegan and vegetarian Pescatarian Fish and poultry Overall</p><p>Sociodemographic</p><p>N (%) 387,731 (94.8%) 7256 (1.8%) 9498 (2.3%) 4625 (1.1%) 409,110</p><p>Age, mean (SD) 56.4 (8.09) 52.9 (7.92) 53.9 (8.03) 56.2 (8.12) 56.3 (8.11)</p><p>Sex, n (%)</p><p>Females 203,550 (52.5%) 4770 (65.7%) 6770 (71.3%) 3477 (75.2%) 218,567 (53.4%)</p><p>Males 184,181 (47.5%) 2486 (34.3%) 2728 (28.7%) 1148 (24.8%) 190,543 (46.6%)</p><p>Deprivation, n (%)</p><p>Lower 133,296 (34.4%) 1882 (25.9%) 2766 (29.1%) 1300 (28.1%) 139,244 (34.0%)</p><p>Middle 130,632 (33.7%) 2337 (32.2%) 3155 (33.2%) 1466 (31.7%) 137,590 (33.6%)</p><p>Higher 123,803 (31.9%) 3037 (41.9%) 3577 (37.7%) 1859 (40.2%) 132,276 (32.3%)</p><p>Ethnicity, n (%)</p><p>White 368,509 (95.0%) 5825 (80.3%) 8882 (93.5%) 4150 (89.7%) 387,366 (94.7%)</p><p>Mixed 5309 (1.4%) 116 (1.6%) 142 (1.5%) 96 (2.1%) 5663 (1.4%)</p><p>South Asian 5776 (1.5%) 1235 (17.0%) 284 (3.0%) 217 (4.7%) 7512 (1.8%)</p><p>Black 5775 (1.5%) 30 (0.4%) 131 (1.4%) 143 (3.1%) 6079 (1.5%)</p><p>Chinese 1206 (0.3%) 9 (0.1%) 10 (0.1%) 5 (0.1%) 1230 (0.3%)</p><p>Any other 1156 (0.3%) 41 (0.6%) 49 (0.5%) 14 (0.3%) 1260 (0.3%)</p><p>Anthropometric</p><p>Height (m), mean (SD) 1.68 (0.09) 1.66 (0.09) 1.67 (0.09) 1.65 (0.09) 1.68 (0.09)</p><p>Weight (kg), mean (SD) 78.4 (15.81) 71.5 (14.66) 70.7 (13.82) 70.3 (14.15) 78.0 (15.82)</p><p>Waist (cm), mean (SD) 90.4 (13.35) 85.1 (12.80) 83.2 (12.07) 83.3 (12.57) 90.1 (13.38)</p><p>Body mass index (kg/m2), mean (SD) 27.4 (4.71) 25.7 (4.63) 25.2 (4.24) 25.6 (4.60) 27.3 (4.72)</p><p>BMI (kg/m2), n (%)</p><p>Underweight 1752 (0.5%) 118 (1.6%) 155 (1.6%) 71 (1.5%) 2096 (0.5%)</p><p>Normal 124,548 (32.1%) 3556 (49.0%) 5010 (52.7%) 2285 (49.4%) 135,399 (33.1%)</p><p>Overweight 167,563 (43.2%) 2531 (34.9%) 3215 (33.8%) 1572 (34.0%) 174,881 (42.7%)</p><p>Obese 93,868 (24.2%) 1051 (14.5%) 1118 (11.8%) 697 (15.1%) 96,734 (23.6%)</p><p>Lifestyle</p><p>Smoking, n (%)</p><p>Never 214,263 (55.3%) 4676 (64.4%) 5437 (57.2%) 2763 (59.7%) 227,139 (55.5%)</p><p>Previous 133,411 (34.4%) 2086 (28.7%) 3390 (35.7%) 1511 (32.7%) 140,398 (34.3%)</p><p>Current 40,057 (10.3%) 494 (6.8%) 671 (7.1%) 351 (7.6%) 41,573 (10.2%)</p><p>Alcohol intake, n (%)</p><p>Daily or almost daily 79,854 (20.6%) 1015 (14.0%) 1825 (19.2%)</p><p>641 (13.9%) 83,335 (20.4%)</p><p>3‑4 times a week 91,813 (23.7%) 1283 (17.7%) 2296 (24.2%) 828 (17.9%) 96,220 (23.5%)</p><p>Once or twice a week 102,119 (26.3%) 1449 (20.0%) 2202 (23.2%) 1063 (23.0%) 106,833 (26.1%)</p><p>1‑3 times a month 43,093 (11.1%) 849 (11.7%) 1103 (11.6%) 522 (11.3%) 45,567 (11.1%)</p><p>Special occasions only 42,704 (11.0%) 1026 (14.1%) 1065 (11.2%) 822 (17.8%) 45,617 (11.2%)</p><p>Never 27,915 (7.2%) 1632 (22.5%) 1004 (10.6%) 743 (16.1%) 31,294 (7.6%)</p><p>Missing 233 (0.1%) 2 (0.0%) 3 (0.0%) 6 (0.1%) 244 (0.1%)</p><p>Sedentary time (h/day), mean (SD) 5.1 (2.26) 4.3 (2.23) 4.3 (2.09) 4.5 (2.33) 5.0 (2.26)</p><p>Physical activity (MET/min/week), mean (SD) 2912.6 (3220.45) 2850.0 (3040.25) 2896.8 (2938.19) 3288.2 (3329.61) 2915.5 (3212.13)</p><p>Health</p><p>Multimorbidity, n (%)</p><p>No 145,488 (37.5%) 3140 (43.3%) 4217 (44.4%) 1787 (38.6%) 154,632 (37.8%)</p><p>Yes 242,243 (62.5%) 4116 (56.7%) 5281 (55.6%) 2838 (61.4%) 254,478 (62.2%)</p><p>Page 6 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>Ta</p><p>bl</p><p>e</p><p>2</p><p>A</p><p>ss</p><p>oc</p><p>ia</p><p>tio</p><p>n</p><p>be</p><p>tw</p><p>ee</p><p>n</p><p>th</p><p>e</p><p>ty</p><p>pe</p><p>s</p><p>of</p><p>d</p><p>ie</p><p>t a</p><p>nd</p><p>c</p><p>an</p><p>ce</p><p>r i</p><p>nc</p><p>id</p><p>en</p><p>ce</p><p>fo</p><p>r m</p><p>od</p><p>el</p><p>s</p><p>0</p><p>an</p><p>d</p><p>4</p><p>Ca</p><p>nc</p><p>er</p><p>s</p><p>ite</p><p>, m</p><p>od</p><p>el</p><p>0</p><p>(u</p><p>na</p><p>dj</p><p>us</p><p>te</p><p>d)</p><p>To</p><p>ta</p><p>l</p><p>Ev</p><p>en</p><p>ts</p><p>M</p><p>ea</p><p>t‑</p><p>ea</p><p>te</p><p>rs</p><p>, r</p><p>ef</p><p>.</p><p>Ev</p><p>en</p><p>ts</p><p>Ve</p><p>ga</p><p>n</p><p>an</p><p>d</p><p>ve</p><p>ge</p><p>ta</p><p>ri</p><p>an</p><p>, H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>Ev</p><p>en</p><p>ts</p><p>Pe</p><p>sc</p><p>at</p><p>ar</p><p>ia</p><p>n,</p><p>H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>Ev</p><p>en</p><p>ts</p><p>Fi</p><p>sh</p><p>a</p><p>nd</p><p>po</p><p>ul</p><p>tr</p><p>y,</p><p>H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>O</p><p>ve</p><p>ra</p><p>ll</p><p>40</p><p>9,</p><p>11</p><p>0</p><p>38</p><p>,0</p><p>42</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>46</p><p>3</p><p>0.</p><p>64</p><p>(0</p><p>.5</p><p>8;</p><p>0</p><p>.7</p><p>0)</p><p>lif</p><p>es</p><p>ty</p><p>le</p><p>fa</p><p>ct</p><p>or</p><p>s</p><p>(s</p><p>m</p><p>ok</p><p>in</p><p>g,</p><p>a</p><p>lc</p><p>oh</p><p>ol</p><p>, t</p><p>ot</p><p>al</p><p>p</p><p>hy</p><p>si</p><p>ca</p><p>l a</p><p>ct</p><p>iv</p><p>ity</p><p>, f</p><p>ru</p><p>its</p><p>, a</p><p>nd</p><p>v</p><p>eg</p><p>et</p><p>ab</p><p>le</p><p>s)</p><p>, c</p><p>om</p><p>or</p><p>bi</p><p>di</p><p>tie</p><p>s</p><p>(p</p><p>re</p><p>se</p><p>nc</p><p>e</p><p>of</p><p>43</p><p>d</p><p>is</p><p>ea</p><p>se</p><p>s)</p><p>, w</p><p>om</p><p>en</p><p>‑r</p><p>ep</p><p>ro</p><p>du</p><p>ct</p><p>iv</p><p>e</p><p>fa</p><p>ct</p><p>or</p><p>s,</p><p>an</p><p>d</p><p>bo</p><p>dy</p><p>m</p><p>as</p><p>s</p><p>in</p><p>de</p><p>x;</p><p>m</p><p>od</p><p>el</p><p>s</p><p>1,</p><p>2</p><p>a</p><p>nd</p><p>3</p><p>a</p><p>re</p><p>p</p><p>re</p><p>se</p><p>nt</p><p>ed</p><p>in</p><p>A</p><p>dd</p><p>iti</p><p>on</p><p>al</p><p>fi</p><p>le</p><p>1</p><p>: T</p><p>ab</p><p>le</p><p>S</p><p>2</p><p>Ta</p><p>bl</p><p>e</p><p>2</p><p>(c</p><p>on</p><p>tin</p><p>ue</p><p>d)</p><p>Ca</p><p>nc</p><p>er</p><p>s</p><p>ite</p><p>, m</p><p>od</p><p>el</p><p>0</p><p>(u</p><p>na</p><p>dj</p><p>us</p><p>te</p><p>d)</p><p>To</p><p>ta</p><p>l</p><p>Ev</p><p>en</p><p>ts</p><p>M</p><p>ea</p><p>t‑</p><p>ea</p><p>te</p><p>rs</p><p>, r</p><p>ef</p><p>.</p><p>Ev</p><p>en</p><p>ts</p><p>Ve</p><p>ga</p><p>n</p><p>an</p><p>d</p><p>ve</p><p>ge</p><p>ta</p><p>ri</p><p>an</p><p>, H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>Ev</p><p>en</p><p>ts</p><p>Pe</p><p>sc</p><p>at</p><p>ar</p><p>ia</p><p>n,</p><p>H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>Ev</p><p>en</p><p>ts</p><p>Fi</p><p>sh</p><p>a</p><p>nd</p><p>po</p><p>ul</p><p>tr</p><p>y,</p><p>H</p><p>R</p><p>95</p><p>%</p><p>C</p><p>I</p><p>P</p><p>va</p><p>lu</p><p>e</p><p>Re</p><p>ct</p><p>um</p><p>40</p><p>9,</p><p>11</p><p>0</p><p>20</p><p>45</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>18</p><p>0.</p><p>72</p><p>(0</p><p>.4</p><p>5;</p><p>1</p><p>.1</p><p>5)</p><p>0.</p><p>16</p><p>3</p><p>35</p><p>0.</p><p>98</p><p>(0</p><p>.7</p><p>0;</p><p>1</p><p>.3</p><p>7)</p><p>0.</p><p>90</p><p>4</p><p>24</p><p>1.</p><p>25</p><p>(0</p><p>.8</p><p>3;</p><p>1</p><p>.8</p><p>6)</p><p>0.</p><p>28</p><p>7</p><p>Pa</p><p>nc</p><p>re</p><p>as</p><p>40</p><p>9,</p><p>11</p><p>0</p><p>11</p><p>33</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>9</p><p>0.</p><p>65</p><p>(0</p><p>.3</p><p>4;</p><p>1</p><p>.2</p><p>6)</p><p>0.</p><p>20</p><p>4</p><p>21</p><p>1.</p><p>08</p><p>(0</p><p>.7</p><p>0;</p><p>1</p><p>.6</p><p>6)</p><p>0.</p><p>74</p><p>1</p><p>6</p><p>0.</p><p>52</p><p>(0</p><p>.2</p><p>3;</p><p>1</p><p>.1</p><p>5)</p><p>0.</p><p>10</p><p>8</p><p>Lu</p><p>ng</p><p>40</p><p>9,</p><p>11</p><p>0</p><p>33</p><p>06</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>27</p><p>0.</p><p>76</p><p>(0</p><p>.5</p><p>2;</p><p>1</p><p>.1</p><p>1)</p><p>0.</p><p>15</p><p>0</p><p>46</p><p>0.</p><p>86</p><p>(0</p><p>.6</p><p>4;</p><p>1</p><p>.1</p><p>5)</p><p>0.</p><p>30</p><p>0</p><p>28</p><p>0.</p><p>81</p><p>(0</p><p>.5</p><p>6;</p><p>1</p><p>.1</p><p>8)</p><p>0.</p><p>28</p><p>2</p><p>M</p><p>el</p><p>an</p><p>om</p><p>a</p><p>40</p><p>9,</p><p>11</p><p>0</p><p>19</p><p>79</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>19</p><p>0.</p><p>68</p><p>(0</p><p>.4</p><p>3;</p><p>1</p><p>.0</p><p>7)</p><p>0.</p><p>09</p><p>6</p><p>28</p><p>0.</p><p>68</p><p>(0</p><p>.4</p><p>7;</p><p>0</p><p>.9</p><p>8)</p><p>0.</p><p>04</p><p>1</p><p>17</p><p>0.</p><p>80</p><p>(0</p><p>.5</p><p>0;</p><p>1</p><p>.2</p><p>9)</p><p>0.</p><p>36</p><p>4</p><p>Br</p><p>ea</p><p>st</p><p>21</p><p>8,</p><p>39</p><p>1</p><p>68</p><p>95</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>13</p><p>8</p><p>0.</p><p>95</p><p>(0</p><p>.8</p><p>0;</p><p>1</p><p>.1</p><p>3)</p><p>0.</p><p>55</p><p>5</p><p>19</p><p>4</p><p>0.</p><p>90</p><p>(0</p><p>.7</p><p>8;</p><p>1</p><p>.0</p><p>4)</p><p>0.</p><p>15</p><p>3</p><p>12</p><p>7</p><p>1.</p><p>11</p><p>(0</p><p>.9</p><p>3;</p><p>1</p><p>.3</p><p>2)</p><p>0.</p><p>24</p><p>3</p><p>Pr</p><p>em</p><p>en</p><p>op</p><p>au</p><p>sa</p><p>l</p><p>54</p><p>,7</p><p>75</p><p>17</p><p>76</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>50</p><p>1.</p><p>00</p><p>(0</p><p>.7</p><p>5;</p><p>1</p><p>.3</p><p>3)</p><p>0.</p><p>97</p><p>0</p><p>52</p><p>0.</p><p>78</p><p>(0</p><p>.5</p><p>9;</p><p>1</p><p>.0</p><p>3)</p><p>0.</p><p>08</p><p>2</p><p>22</p><p>0.</p><p>91</p><p>(0</p><p>.6</p><p>0;</p><p>1</p><p>.3</p><p>9)</p><p>0.</p><p>65</p><p>7</p><p>Po</p><p>st</p><p>m</p><p>en</p><p>op</p><p>au</p><p>sa</p><p>l</p><p>13</p><p>0,</p><p>62</p><p>5</p><p>36</p><p>68</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>71</p><p>0.</p><p>92</p><p>(0</p><p>.7</p><p>3;</p><p>1</p><p>.1</p><p>6)</p><p>0.</p><p>30</p><p>0</p><p>11</p><p>3</p><p>0.</p><p>91</p><p>(0</p><p>.7</p><p>5;</p><p>1</p><p>.1</p><p>0)</p><p>0.</p><p>18</p><p>6</p><p>79</p><p>1.</p><p>06</p><p>(0</p><p>.8</p><p>5;</p><p>1</p><p>.3</p><p>3)</p><p>0.</p><p>60</p><p>4</p><p>U</p><p>te</p><p>rin</p><p>e</p><p>21</p><p>8,</p><p>39</p><p>1</p><p>11</p><p>31</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>23</p><p>1.</p><p>15</p><p>(0</p><p>.7</p><p>6;</p><p>1</p><p>.7</p><p>4)</p><p>0.</p><p>52</p><p>2</p><p>30</p><p>1.</p><p>09</p><p>(0</p><p>.7</p><p>5;</p><p>1</p><p>.5</p><p>6)</p><p>0.</p><p>65</p><p>5</p><p>18</p><p>1.</p><p>07</p><p>(0</p><p>.6</p><p>7;</p><p>1</p><p>.7</p><p>1)</p><p>0.</p><p>76</p><p>4</p><p>O</p><p>va</p><p>ry</p><p>21</p><p>8,</p><p>39</p><p>1</p><p>87</p><p>0</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>19</p><p>1.</p><p>14</p><p>(0</p><p>.7</p><p>2;</p><p>1</p><p>.8</p><p>0)</p><p>0.</p><p>58</p><p>3</p><p>28</p><p>1.</p><p>07</p><p>(0</p><p>.7</p><p>3;</p><p>1</p><p>.5</p><p>7)</p><p>0.</p><p>71</p><p>6</p><p>18</p><p>1.</p><p>20</p><p>(0</p><p>.7</p><p>5;</p><p>1</p><p>.9</p><p>1)</p><p>0.</p><p>45</p><p>5</p><p>Pr</p><p>os</p><p>ta</p><p>te</p><p>19</p><p>0,</p><p>54</p><p>3</p><p>74</p><p>92</p><p>1.</p><p>00</p><p>(r</p><p>ef</p><p>.)</p><p>47</p><p>0.</p><p>57</p><p>(0</p><p>.4</p><p>3;</p><p>0</p><p>.7</p><p>6)</p><p>al. BMC Medicine (2022) 20:79</p><p>month) were excluded (Orlish et al [33], Tantamango</p><p>et  al., [15] and Penniecook et  al [34]) the associa-</p><p>tion of vegetarians with overall cancer remained</p><p>significant (RR: ;0.89, 95% CI: 0.84; 0.94), while the</p><p>association for colorectal cancer remained no signifi-</p><p>cant (Additional file 1: Fig S7).</p><p>For pescatarian diets, there were 3 studies (1482</p><p>events) for overall cancer [13, 15], 4 studies (167 events)</p><p>for colorectal cancer [13, 31, 33], 3 studies (61 events) for</p><p>Fig. 4 Forest plot of prospective cohort studies evaluating the summary risk ratios of colon and rectal cancer of vegetarians and pescatarians</p><p>compared with meat‑eaters (reference). RR, risk ratio; CI, confidence interval</p><p>Page 12 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>lung cancer [13, 32], 4 studies (250 events) for prostate</p><p>cancer [13, 32, 35] and 5 studies (585 events) for breast</p><p>cancer [13, 17, 32, 34] (Fig.  3). Compared with meat-</p><p>eaters, pescatarians had a 9% lower risk of overall cancer</p><p>(pooled RR: 0.91, 95% CI 0.86; 0.96) and a 15% lower risk</p><p>for colorectal cancer (RR: 0.75, 95% CI: 0.59; 0.94), but no</p><p>significant associations were observed for lung (RR 0.75,</p><p>95% CI 0.54; 1.05), prostate (RR 0.97, 95% CI 0.76; 1.23)</p><p>or breast cancer (RR: 0.95, 95% CI: 0.82; 1.10) (Fig.  3).</p><p>When the pooled data were examined for colon and rec-</p><p>tal cancer separately, there were no differences in the</p><p>risk for vegetarians or pescatarians compared with meat-</p><p>eaters (Fig.  4). Similarly, when breast cancer was strati-</p><p>fied by menopausal status, the risk did not differ between</p><p>vegetarians, pescatarians and meat-eaters (Fig. 5). Similar</p><p>results were observed when the study of Penniecook et al</p><p>[34]) was removed from the the breast cancer analysis on</p><p>vegetarians, as their definition of vegetarians included a</p><p>Fig. 5 Forest plot of prospective cohort studies evaluating summary risk ratios of breast cancer in premenopausal and postmenopausal women for</p><p>vegetarians and pescatarians compared with meat‑eaters (reference). RR, risk ratio; CI, confidence interval</p><p>Page 13 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>low meat consumption (less than once a month) (Addi-</p><p>tional file  1: Fig S8) Finally, when Orlish et  al [33] was</p><p>removed from the analysis for colorectal cancer as the</p><p>authors included low meat consumption (less than once</p><p>a month) on their definition of pescatarian, the associa-</p><p>tion became non significant (RR: 0.81, 95% CI: 0.65; 1.01)</p><p>(Additional file 1: Fig S9).</p><p>Discussion</p><p>This analysis of the UK Biobank cohort suggests that,</p><p>compared with meat-eaters, vegetarians had a lower</p><p>risk of all cancers combined and of colorectal (especially</p><p>colon and proximal colon) and prostate cancer and that</p><p>pescatarians had a lower risk of all cancers and mela-</p><p>noma. These associations were independent of major</p><p>confounding factors, including socio-demographics,</p><p>lifestyle, multimorbidity and adiposity. When the results</p><p>were pooled with eight previous studies in a meta-anal-</p><p>ysis, the associations with overall and colorectal cancers</p><p>persisted, but the results relating to other specific cancer</p><p>sites were inconclusive.</p><p>Although evidence for associations between red and</p><p>processed meat consumption and increased risk of can-</p><p>cer has been widely reported [6, 36–39], there is lim-</p><p>ited and equivocal evidence for alternative diets, such</p><p>as vegetarian or pescatarian, on cancer risk [18, 19,</p><p>40]. In our study, vegetarians and pescatarians had a</p><p>10% and 9% lower risk of overall cancer compared with</p><p>meat-eaters, respectively. These findings of lower risk</p><p>agree with evidence reported in previous studies [18].</p><p>A meta-analysis published in 2017, which included</p><p>38,033 participants from two prospective cohort stud-</p><p>ies from the USA and the UK, of whom 1976 developed</p><p>incident all-cause cancers, showed that vegetarians had</p><p>an 8% lower risk [18]. However, an earlier meta-analysis</p><p>by Huang et  al. in 2012, which included 124,706 par-</p><p>ticipants from the UK, the Netherlands and Germany,</p><p>reported a larger reduction of 18% for overall cancer</p><p>incidence in vegetarians compared with meat-eaters</p><p>[41].</p><p>Regarding specific cancers, in a study including 77,659</p><p>participants who were followed up for 7.3 years, Orlich</p><p>et  al. reported that vegetarians had a 22% lower inci-</p><p>dence risk of colorectal cancer (HR 0.78 [95% CI, 0.64;</p><p>0.95] (490 cases)) compared with meat-eaters. The reduc-</p><p>tion in risk was stronger for pescatarians, who had a 43%</p><p>lower risk of colorectal cancer (HR 0.57 [95% CI, 0.40;</p><p>0.82]) [33]. In a relatively recent meta-analysis, Godos</p><p>et al. [19], which included a total of nine studies includ-</p><p>ing six cohorts accounting for 686,629 individuals, and</p><p>3441, 4062 and 1935 cases of breast, colorectal and pros-</p><p>tate cancer, respectively, reported no significant differ-</p><p>ences between vegetarians and non-vegetarians in the</p><p>risk of breast and prostate cancer. However, they did find</p><p>a lower risk of colorectal cancer in semi-vegetarians (RR</p><p>0.86 [95% CI 0.79; 0.94]) and pescatarians (RR 0.67 [95%</p><p>CI 0.53; 0.83]) [19]. In our study using UK Biobank data,</p><p>we found a 27% lower risk of colorectal cancer in veg-</p><p>etarians. When the analyses were stratified by subtype,</p><p>cancers in the proximal colon and colon showed a 57%</p><p>and 31% lower risk in vegetarians, respectively, with no</p><p>associations observed for distal colon or rectal cancer.</p><p>Interestingly, for pescatarians, we observed significant</p><p>associations for colorectal and colon cancer only in the</p><p>minimally adjusted model, and these associations were</p><p>completely attenuated when the analyses were adjusted</p><p>for socio-demographics, lifestyle and BMI. Similar to</p><p>breast and lung cancers, the associations with diet type</p><p>were significant in the unadjusted model but disappeared</p><p>completely in the most adjusted model. Indeed, these</p><p>results are in agreement with previous studies, espe-</p><p>cially for breast cancer, where associations were attenu-</p><p>ated after accounting for BMI [13, 17, 32, 34]. No studies</p><p>have reported significant differences in lung cancer risk</p><p>for vegetarians or pescatarians versus meat-eaters in</p><p>maximally adjusted models [13, 17, 32]. Our finding of no</p><p>significant associations with site-specific cancers among</p><p>poultry eaters was similar to those reported by Cade et al.</p><p>in 2010 [17].</p><p>Some of the mechanisms that could explain the asso-</p><p>ciations between vegetarian diet and cancer risk are the</p><p>presence of bioactive compounds in plant-based diets,</p><p>such as fibre, phenol, polyphenol, and sulphuric com-</p><p>pounds, and other antioxidants compounds, including</p><p>vitamins. These compounds have been shown to have</p><p>anti-carcinogenic effects in experimental models and</p><p>epidemiological studies [42, 43]. On the other hand,</p><p>this dietary pattern has some deficient intake of certain</p><p>nutrients such as iron and vitamin B12. For instance,</p><p>long-chain n-3 fatty acids, such as eicosapentaenoic acid</p><p>and docosahexaenoic acid, are lower in vegetarians [44].</p><p>Therefore, decreased intakes of some of these nutrients</p><p>have been related to a higher cancer incidence in some</p><p>studies [45]. We cannot discard that some of the associa-</p><p>tions described may be mediated through other risk fac-</p><p>tors such as adiposity [46–48] or smoking [48].</p><p>It is important to notice that not all vegetarians’ diets</p><p>are healthy [49]; higher consumption of ultra-processed</p><p>food could reduce the benefit of a vegetarian diet on</p><p>cancer risk. In our sample, vegetarian and pescatarians</p><p>reported eating more crisps and pizza than meat-eaters.</p><p>Despite low/moderate heterogeneity in our meta-anal-</p><p>ysis, studies differed in a variety of ways that could lead</p><p>to inconsistencies in their findings. Some of these include</p><p>the length of follow-up (ranging from 4.1 to 20.3 years),</p><p>differences in confounding factors controlled for in each</p><p>Page 14 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>study, differences in sample size across cohorts</p><p>(rang-</p><p>ing from 10,210 to 409,110) and risks of bias attribut-</p><p>able to the design of the studies which varied from low</p><p>to moderate [18]. Studies also differed in how vegetarian</p><p>and other types of diets were defined and for how long</p><p>participants had been following their attributed type of</p><p>diet [41]. However, the measured heterogeneity between</p><p>studies included in the meta-analyses was low.</p><p>Strength and limitations</p><p>UK Biobank is a large, prospective, general population</p><p>cohort with data on diet and a wide range of poten-</p><p>tial confounders and health outcomes. However, the</p><p>UK Biobank is not representative of the general UK</p><p>population regarding lifestyle and baseline health [50].</p><p>Moreover, the participants who have full data available</p><p>and therefore were included in this study were leaner</p><p>and more affluent than those excluded due to missing</p><p>data and prevalent cancer at baseline (Additional file  1:</p><p>Table  S3). Regardless of these differences, our risk esti-</p><p>mates were similar to other more population-represent-</p><p>ative cohorts [20]. In the UK Biobank, dietary data were</p><p>collected from all participants on one occasion; therefore,</p><p>we cannot rule out that the type of diet reported was not</p><p>modified over the length of the study. However, an analy-</p><p>sis of the repeatability of the touchscreen questionnaire</p><p>in a sub-set of participants (n = 20.348), who repeated</p><p>the assessment centre visit approximately 4 years after</p><p>recruitment, showed that the dietary touchscreen vari-</p><p>ables, available for the full cohort, reliably rank partici-</p><p>pants according to intakes of the main food groups. In</p><p>our study, we were unable to investigate the association of</p><p>vegan diet with cancer risk because there were very small</p><p>numbers (n = 57) of participants who were vegan. Also,</p><p>we were not able to include energy intake as a covariate</p><p>as data were available for only half of the cohort popula-</p><p>tion. Moreover, exposure specificity is also a limitation as</p><p>there was a lack of data on what alternative nutrients or</p><p>food sources were used to replace meat or fish consump-</p><p>tion. An additional limitation of the UK Biobank findings</p><p>is that data were collected at a single time point resulting</p><p>in the inability to properly adjust for changes in the expo-</p><p>sure or covariates over time.</p><p>For the meta-analysis, some data limitations should</p><p>be considered when interpreting the results. The only</p><p>available evidence was from high-income countries in</p><p>Europe and North America, i.e. the USA, Canada, the UK</p><p>and the Netherlands. Therefore, associations between</p><p>dietary patterns and cancer in other continents and in</p><p>low-income and middle-income countries remain to be</p><p>investigated. The number of available studies was gen-</p><p>erally low for the cancer-specific analyses. This limited</p><p>the possibility of exploring subgroup analysis, but the</p><p>heterogeneity across studies was relativity low. The defi-</p><p>nitions used for different types of diets differed between</p><p>studies, which could introduce bias and reduce the like-</p><p>lihood of detecting associations. Future research in this</p><p>area should aim to standardize diet classifications relat-</p><p>ing to the types of food consumed and their frequency of</p><p>consumption.</p><p>Conclusion</p><p>Our UK Biobank findings suggest that, compared with</p><p>meat-eaters, vegetarians had a lower risk of cancer over-</p><p>all, probably due to a lower risk of colorectal and pros-</p><p>tate cancer. Pescatarians also had a lower overall risk of</p><p>cancer, but the relationships with specific contributory</p><p>cancers were unclear. However, when our risk estimates</p><p>were pooled with those from previous prospective cohort</p><p>studies, there was no conclusive evidence in relation to</p><p>site-specific cancers, even though the associations with</p><p>overall cancer risk were significant. Larger studies with</p><p>longer follow-up periods and better classification of diet</p><p>types are needed to elucidate the benefits, or otherwise,</p><p>of vegetarian and pescatarian diets on the risk of individ-</p><p>ual cancers.</p><p>Abbreviations</p><p>BMI: Body mass index; CI: Confidence interval; HR: Hazard ratio; MET: Metabolic</p><p>equivalent; n: Number; PA: Physical activity; RR: Relative risk; SD: Standard</p><p>deviation.</p><p>Supplementary Information</p><p>The online version contains supplementary material available at https:// doi.</p><p>org/ 10. 1186/ s12916‑ 022‑ 02257‑9.</p><p>Additional file 1: Methods. Fig. S1. Flowchart of UK Biobank participants.</p><p>Table S1. Criteria for the Newcastle‑Ottawa Scale regarding star allocation</p><p>to assess quality of studies. Table S2. Details of dietary characteristics for</p><p>175,499 UK Biobank participants with data available for the 24 hrs diet</p><p>recall. Table S3. Baseline characteristics people with missing data who</p><p>were not included in the study. Table S4. Association between types of</p><p>diet and cancer incidence for Models 1‑3. Table S5. Landmark sensitivity</p><p>analysis: Association between types of diet and cancer incidence for all</p><p>Models after excluding events in the first 2 years of follow‑up. Table S6.</p><p>Characteristics of the cohorts included in systematic review. Table S7.</p><p>Definitions of vegetarian, lacto‑ovo‑vegetarian, pescatarian and poultry</p><p>diets used in the various studies. Table S8. Quality assessment of studies</p><p>using a modified Newcastle‑Ottawa scale for assessing studies in the</p><p>systematic review of vegetarian diet and cancer risk. Table S9. Risk of bias</p><p>assessment. Fig. S2. Funnel plot of prospective cohort studies evaluating</p><p>summary hazard ratios of colorectal, lung, prostate, breast and overall</p><p>cancer for vegetarians versus meat‑eaters. Fig. S3. Funnel plot of prospec‑</p><p>tive cohort studies evaluating summary hazard ratios of colorectal, lung,</p><p>prostate, breast and overall cancer for pescatarians versus meat‑eaters.</p><p>Fig. S4. Funnel plot of prospective cohort studies evaluating summary</p><p>hazard ratios of colon and rectum cancer for vegetarian and pescatarians</p><p>versus meat‑eaters. Fig. S5. Funnel plot of prospective cohort studies</p><p>evaluating summary hazard ratios of breast cancer in premenopausal and</p><p>postmenopausal for vegetarian and pescatarians versus meat‑eaters. Fig</p><p>S6. Sensitivity analysis of prospective cohort studies evaluating summary</p><p>risk ratios of of Lacto‑Ovo vegetarians defined by Tantamango et al. for</p><p>https://doi.org/10.1186/s12916-022-02257-9</p><p>https://doi.org/10.1186/s12916-022-02257-9</p><p>Page 15 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>overall cancer, Orlish et al. for colorectal cancer, and Penniecook et al for</p><p>breast cancer, compared with non‑vegetarians (reference). Fig S7. Sensi‑</p><p>tivity analysis of prospective cohort studies evaluating summary risk ratios</p><p>of overall, colorectal, lung, prostate and breast cancer for vegetarians com‑</p><p>pared with meat‑eaters. Fig S8. Sensitivity analysis of prospective cohort</p><p>studies evaluating summary risk ratios of breast cancer in premenopausal</p><p>and postmenopausal women for vegetarians compared with meat‑eaters</p><p>(reference). Fig S9. Sensitivity analysis of prospective cohort studies evalu‑</p><p>ating summary risk ratios of colorectal, lung, prostate, breast and overall</p><p>cancer of pescatarians compared with meat‑eaters.</p><p>Data share statement</p><p>Data described in the manuscript, codebook and analytic code will be made</p><p>available upon request pending approval from the UK Biobank management</p><p>team.</p><p>Authors’ contributions</p><p>SP‑S, DA, FH, JPP and CC‑M contributed to the conception and design of</p><p>the study and advised on all statistical aspects. SPS conducted the statistical</p><p>analysis assisted by CC‑M and FH. All authors interpreted the data. SPS, DA</p><p>and CC‑M drafted the first draft. All authors reviewed the manuscript and</p><p>approved the final version to be published. SP‑S, CCM, FH and JPP had full</p><p>access to all the data in the study and take responsibility for the integrity</p><p>of the data and the accuracy of the data analysis. CCM, FH and JPP are the</p><p>guarantors. The corresponding author attests that all listed authors meet the</p><p>authorship criteria and that no others meeting the criteria have been omitted.</p><p>All authors</p><p>read and approved the final version of the manuscript.</p><p>Funding</p><p>The UK Biobank was supported by the Wellcome Trust, Medical Research</p><p>Council, Department of Health, Scottish Government and Northwest Regional</p><p>Development Agency. It also had funding from the Welsh Assembly Govern‑</p><p>ment and British Heart Foundation. S.P.‑S. and FP‑R receive financial support</p><p>from the Chilean government for doing their PhD (ANID‑Becas Chile). KML is</p><p>supported by a National Health and Medical Research Council Emerging Lead‑</p><p>ership Fellowship [APP1173803]. This research was also funded by grant num‑</p><p>ber IIG_FULL_2020_032 from the Wereld Kanker Onderzoek Fonds (WKOF), as</p><p>part of the World Cancer Research Fund International grant programme. The</p><p>research was designed, conducted, analysed and interpreted by the authors</p><p>entirely independently of the funding sources.</p><p>Availability of data and materials</p><p>This research has been conducted using the UK Biobank Resource under</p><p>Application Number 7155. The UK Biobank is an open‑access resource, and</p><p>all bona fide researchers can use it for approved research by registering and</p><p>applying at http:// www. ukbio bank. ac. uk/ regis ter‑ apply/.</p><p>Declarations</p><p>Ethics approval and consent to participate</p><p>All participants provided written informed consent before enrolment in the</p><p>UK Biobank, which was conducted in accordance with the Declaration of</p><p>Helsinki. The UK Biobank study, and the sharing of anonymized data with the</p><p>research community, was approved by the North West Multi‑center Research</p><p>Ethics Committee (REC reference: 12/NW/03820).</p><p>Consent for publication</p><p>Not applicable.</p><p>Competing interests</p><p>The authors declare that they have no competing interests.</p><p>Author details</p><p>1 Institute of Health and Wellbeing, University of Glasgow, Glasgow G12</p><p>8RZ, UK. 2 British Heart Foundation Cardiovascular Research Centre, Institute</p><p>of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow G12</p><p>8TA, UK. 3 Department of Process and Evaluation, Faculty of Health and Life</p><p>Sciences, Universidad Católica de Temuco, Temuco, Chile. 4 Facultad de</p><p>Medicina, Universidad Diego Portales, Santiago, Chile. 5 Department of Applied</p><p>Sciences, Faculty of Health and Life Sciences, Northumbria University, EBD223</p><p>Ellison Building, Newcastle upon Tyne NE1 8ST, UK. 6 Newcastle University</p><p>Centre for Cancer, Population Health Sciences Institute, Newcastle Univer‑</p><p>sity, Newcastle upon Tyne, UK. 7 Human Nutrition Research Centre, Centre</p><p>for Healthier Lives, Population Health Sciences Institute, Newcastle University,</p><p>Newcastle upon Tyne NE2 4HH, UK. 8 Institute for Physical Activity and Nutri‑</p><p>tion, School of Exercise and Nutrition Sciences, Deakin University, Geelong,</p><p>VIC 3220, Australia. 9 Education, Physical Activity and Health Research Unit,</p><p>University Católica del Maule, 3466706 Talca, Chile.</p><p>Received: 28 September 2021 Accepted: 13 January 2022</p><p>Published: 24 February 2022</p><p>References</p><p>1. Collaborators GBDD. Health effects of dietary risks in 195 countries,</p><p>1990‑2017: a systematic analysis for the Global Burden of Disease</p><p>Study 2017. Lancet. 2019;393(10184):1958–72.</p><p>2. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A,</p><p>et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence</p><p>and mortality worldwide for 36 cancers in 185 countries. CA Cancer J</p><p>Clin. 2021;71(3):209–49.</p><p>3. Key TJ, Bradbury KE, Perez‑Cornago A, Sinha R, Tsilidis KK, Tsugane S.</p><p>Diet, nutrition, and cancer risk: what do we know and what is the way</p><p>forward? BMJ. 2020;368:m511.</p><p>4. Negri E, Franceschi S, Parpinel M, La Vecchia C. Fiber intake and risk of</p><p>colorectal cancer. Cancer Epidemiol Prev Biomarkers. 1998;7(8):667–71.</p><p>5. Franceschi S, Favero A, LaVecchia C, Negri E, Conti E, Montella M,</p><p>et al. Food groups and risk of colorectal cancer in Italy. Int J Cancer.</p><p>1997;72(1):56–61.</p><p>6. Johnston BC, Zeraatkar D, Han MA, Vernooij RWM, Valli C, El Dib R, et al.</p><p>Unprocessed red meat and processed meat consumption: dietary</p><p>guideline recommendations from the Nutritional Recommendations</p><p>(NutriRECS) Consortium. Ann Intern Med. 2019;171(10):756–64.</p><p>7. Rosenfeld DL, Burrow AL. Vegetarian on purpose: understanding the</p><p>motivations of plant‑based dieters. Appetite. 2017;116:456–63.</p><p>8. Leitzmann C. Vegetarian nutrition: past, present, future. Am J Clin Nutr.</p><p>2014;100(Suppl 1):496S–502S.</p><p>9. Key TJ, Appleby PN, Rosell MS. Health effects of vegetarian and vegan</p><p>diets. Proc Nutr Soc. 2006;65(1):35–41.</p><p>10. How many adults in the U.S. are vegetarian and vegan? How many adults</p><p>eat vegetarian and vegan meals when eating out? https:// www. vrg. org/</p><p>nutsh ell/ Polls/ 2019_ adults_ veg. htm.</p><p>11. Appleby PN, Crowe FL, Bradbury KE, Travis RC, Key TJ. Mortality in vegetar‑</p><p>ians and comparable nonvegetarians in the United Kingdom. Am J Clin</p><p>Nutr. 2016;103(1):218–30.</p><p>12. Key TJ, Appleby PN, Spencer EA, Travis RC, Roddam AW, Allen NE.</p><p>Mortality in British vegetarians: results from the European Prospective</p><p>Investigation into Cancer and Nutrition (EPIC‑Oxford). Am J Clin Nutr.</p><p>2009;89(5):1613s–9s.</p><p>13. Key TJ, Appleby PN, Crowe FL, Bradbury KE, Schmidt JA, Travis RC. Cancer</p><p>in British vegetarians: updated analyses of 4998 incident cancers in a</p><p>cohort of 32,491 meat eaters, 8612 fish eaters, 18,298 vegetarians, and</p><p>2246 vegans. Am J Clin Nutr. 2014;100(Suppl 1):378S–85S.</p><p>14. Orlich MJ, Singh PN, Sabate J, Jaceldo‑Siegl K, Fan J, Knutsen S, et al. Veg‑</p><p>etarian dietary patterns and mortality in Adventist Health Study 2. JAMA</p><p>Intern Med. 2013;173(13):1230–8.</p><p>15. Tantamango‑Bartley Y, Jaceldo‑Siegl K, Fan J, Fraser G. Vegetarian diets</p><p>and the incidence of cancer in a low‑risk population. Cancer Epidemiol</p><p>Biomarkers Prev. 2013;22(2):286–94.</p><p>16. Key TJ, Appleby PN, Spencer EA, Travis RC, Allen NE, Thorogood M, et al.</p><p>Cancer incidence in British vegetarians. Br J Cancer. 2009;101(1):192–7.</p><p>17. Cade JE, Taylor EF, Burley VJ, Greenwood DC. Common dietary patterns</p><p>and risk of breast cancer: analysis from the United Kingdom Women’s</p><p>Cohort Study. Nutr Cancer. 2010;62(3):300–6.</p><p>18. Dinu M, Abbate R, Gensini GF, Casini A, Sofi F. Vegetarian, vegan diets</p><p>and multiple health outcomes: a systematic review with meta‑analysis of</p><p>observational studies. Crit Rev Food Sci. 2017;57(17):3640–9.</p><p>http://www.ukbiobank.ac.uk/register-apply/</p><p>https://www.vrg.org/nutshell/Polls/2019_adults_veg.htm</p><p>https://www.vrg.org/nutshell/Polls/2019_adults_veg.htm</p><p>Page 16 of 16Parra‑Soto et al. BMC Medicine (2022) 20:79</p><p>•</p><p>fast, convenient online submission</p><p>•</p><p>thorough peer review by experienced researchers in your field</p><p>•</p><p>rapid publication on acceptance</p><p>•</p><p>support for research data, including large and complex data types</p><p>•</p><p>gold Open Access which fosters wider collaboration and increased citations</p><p>maximum visibility for your research: over 100M website views per year •</p><p>At BMC, research is always in progress.</p><p>Learn more biomedcentral.com/submissions</p><p>Ready to submit your researchReady to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from:</p><p>19. Godos J, Bella F, Sciacca S, Galvano F, Grosso G. Vegetarianism and breast,</p><p>colorectal and prostate cancer risk: an overview and meta‑analysis of</p><p>cohort studies. J Hum Nutr Diet. 2017;30(3):349–59.</p><p>20. Collins R. What makes UK Biobank special? Lancet. 2012;379(9822):1173–4.</p><p>21. Palmer LJ. UK Biobank: bank on it. Lancet. 2007;369(9578):1980–2.</p><p>22. Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK</p><p>biobank: an open access resource for identifying the causes of a</p><p>wide range of complex diseases of middle and old age. PLoS Med.</p><p>2015;12(3):e1001779.</p><p>23. Welcome to UK Biobank. http:// www. ukbio bank. ac. uk.</p><p>24. Petermann‑Rocha F, Parra‑Soto S, Gray S, Anderson J, Welsh P, Gill J, et al.</p><p>Vegetarians, fish, poultry, and meat‑eaters: who has higher risk of cardio‑</p><p>vascular disease incidence and mortality? A prospective study from UK</p><p>Biobank. Eur Heart J. 2021;42(12):1136–43.</p><p>25. Galante J, Adamska L, Young A, Young H, Littlejohns TJ, Gallacher J, et al.</p><p>The acceptability of repeat Internet‑based hybrid diet assessment of</p><p>previous 24‑h dietary intake: administration of the Oxford WebQ in UK</p><p>Biobank. Br J Nutr. 2016;115(4):681–6.</p><p>26. Bradbury KE, Young HJ, Guo W, Key TJ. Dietary assessment in UK Biobank:</p><p>an evaluation of the performance of the touchscreen dietary question‑</p><p>naire. J Nutr Sci. 2018;7:e6.</p><p>27. Version 2 April Internet. www. ipaq. ki. se.</p><p>28. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD,</p><p>et al. The PRISMA 2020 statement: an updated guideline for reporting</p><p>systematic reviews. BMJ. 2021;372:n71.</p><p>29. Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan‑a web and</p><p>mobile app for systematic reviews. Syst Rev‑London. 2016;5(1):1–0.</p><p>30. Stang A. Critical evaluation of the Newcastle‑Ottawa scale for the assess‑</p><p>ment of the quality of nonrandomized studies in meta‑analyses. Eur J</p><p>Epidemiol. 2010;25(9):603–5.</p><p>31. Gilsing AM, Schouten LJ, Goldbohm RA, Dagnelie PC, van den Brandt PA, Wei‑</p><p>jenberg MP. Vegetarianism, low meat consumption and the risk of colorectal</p><p>cancer in a population based cohort study. Sci Rep. 2015;5:13484.</p><p>32. Gilsing AM, Weijenberg MP, Goldbohm RA, Dagnelie PC, van den Brandt</p><p>PA, Schouten LJ. Vegetarianism, low meat consumption and the risk of</p><p>lung, postmenopausal breast and prostate cancer in a population‑based</p><p>cohort study. Eur J Clin Nutr. 2016;70(6):723–9.</p><p>33. Orlich MJ, Singh PN, Sabate J, Fan J, Sveen L, Bennett H, et al. Vegetarian</p><p>dietary patterns and the risk of colorectal cancers. JAMA Intern Med.</p><p>2015;175(5):767–76.</p><p>34. Penniecook‑Sawyers JA, Jaceldo‑Siegl K, Fan J, Beeson L, Knutsen S, Her‑</p><p>ring P, et al. Vegetarian dietary patterns and the risk of breast cancer in a</p><p>low‑risk population. Br J Nutr. 2016;115(10):1790–7.</p><p>35. Tantamango‑Bartley Y, Knutsen SF, Knutsen R, Jacobsen BK, Fan J, Beeson</p><p>WL, et al. Are strict vegetarians protected against prostate cancer? Am J</p><p>Clin Nutr. 2016;103(1):153–60.</p><p>36. Anderson JJ, Darwis NDM, Mackay DF, Celis‑Morales CA, Lyall DM, Sattar</p><p>N, et al. Red and processed meat consumption and breast cancer: UK</p><p>Biobank cohort study and meta‑analysis. Eur J Cancer. 2018;90:73–82.</p><p>37. Bradbury KE, Murphy N, Key TJ. Diet and colorectal cancer in UK Biobank:</p><p>a prospective study. Int J Epidemiol. 2020;49(1):246–58.</p><p>38. Handel MN, Cardoso I, Rasmussen KM, Rohde JF, Jacobsen R, Nielsen SM,</p><p>et al. Processed meat intake and chronic disease morbidity and mortal‑</p><p>ity: an overview of systematic reviews and meta‑analyses. PLoS One.</p><p>2019;14(10):e0223883.</p><p>39. World Cancer Research Fund/American Institute for Cancer Research.</p><p>Continuous Update Project Expert Report 2018. Diet, nutrition, physical</p><p>activity and cancer: a global perspective. About the Third Expert Report.</p><p>2018.</p><p>40. Veettil SK, Wong TY, Loo YS, Playdon MC, Lai NM, Giovannucci EL,</p><p>et al. Role of diet in colorectal cancer incidence: umbrella review of</p><p>meta‑analyses of prospective observational studies. JAMA Netw Open.</p><p>2021;4(2):e2037341.</p><p>41. Huang T, Yang B, Zheng JS, Li GP, Wahlqvist ML, Li D. Cardiovascular dis‑</p><p>ease mortality and cancer incidence in vegetarians: a meta‑analysis and</p><p>systematic review. Ann Nutr Metab. 2012;60(4):233–40.</p><p>42. Subramaniam S, Selvaduray KR, Radhakrishnan AK. Bioactive compounds:</p><p>natural defense against cancer? Biomolecules. 2019;9(12):758.</p><p>43. Silveira DSC, Veronez LC, Lopes‑Junior LC, Anatriello E, Brunaldi MO,</p><p>Pereira‑da‑Silva G. Lactobacillus bulgaricus inhibits colitis‑associated</p><p>cancer via a negative regulation of intestinal inflammation in</p><p>azoxymethane/dextran sodium sulfate model. World J Gastroenterol.</p><p>2020;26(43):6782–94.</p><p>44. Burns‑Whitmore B, Froyen E, Heskey C, Parker T, San Pablo G. Alpha‑</p><p>linolenic and linoleic fatty acids in the vegan diet: do they require dietary</p><p>reference intake/adequate intake special consideration? Nutrients.</p><p>2019;11(10):2365.</p><p>45. Siriwardhana N, Kalupahana NS, Moustaid‑Moussa N. Health benefits of</p><p>n‑3 polyunsaturated fatty acids: eicosapentaenoic acid and docosahexae‑</p><p>noic acid. Adv Food Nutr Res. 2012;65:211–22.</p><p>46. Arnold M, Leitzmann M, Freisling H, Bray F, Romieu I, Renehan A, et al.</p><p>Obesity and cancer: an update of the global impact. Cancer Epidemiol.</p><p>2016;41:8–15.</p><p>47. Arnold M, Pandeya N, Byrnes G, Renehan AG, Stevens GA, Ezzati M, et al.</p><p>Global burden of cancer attributable to high body‑mass index in 2012: a</p><p>population‑based study. Lancet Oncol. 2015;16(1):36–46.</p><p>48. Brown KF, Rumgay H, Dunlop C, Ryan M, Quartly F, Cox A, et al. The frac‑</p><p>tion of cancer attributable to modifiable risk factors in England, Wales,</p><p>Scotland, Northern Ireland, and the United Kingdom in 2015. Br J Cancer.</p><p>2018;118(8):1130–41.</p><p>49. Gehring J, Touvier M, Baudry J, Julia C, Buscail C, Srour B, et al. Consump‑</p><p>tion of ultra‑processed foods by pesco‑vegetarians, vegetarians, and</p><p>vegans: associations with duration and age at diet initiation. J Nutr.</p><p>2021;151(1):120–31.</p><p>50. Batty GD, Gale CR, Kivimäki M, Deary IJ, Bell S. Comparison of risk factor</p><p>associations in UK Biobank against representative, general population</p><p>based studies with conventional response rates: prospective cohort</p><p>study and individual participant meta‑analysis. BMJ. 2020;368:m131.</p><p>Publisher’s Note</p><p>Springer Nature remains neutral with regard to jurisdictional claims in pub‑</p><p>lished maps and institutional affiliations.</p><p>http://www.ukbiobank.ac.uk</p><p>http://www.ipaq.ki.se</p><p>Association of meat, vegetarian, pescatarian and fish-poultry diets with risk of 19 cancer sites and all cancer: findings from the UK Biobank prospective cohort study and meta-analysis</p><p>Abstract</p><p>Background:</p><p>Methods:</p><p>Results:</p><p>Conclusions:</p><p>Introduction</p><p>Methods</p><p>UK Biobank</p><p>Outcomes</p><p>Types of diets</p><p>Covariates</p><p>Statistical analyses</p><p>Meta-analysis</p><p>Results</p><p>UK Biobank</p><p>Meta-analysis</p><p>Discussion</p><p>Strength and limitations</p><p>Conclusion</p><p>Data share statement</p><p>References</p>